Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change

TL;DR

This paper introduces BetaRCE, a novel method for generating counterfactual explanations with probabilistic guarantees of robustness to model changes, applicable to any model and change type, improving explanation stability and interpretability.

Contribution

It proposes a theoretical framework for probabilistic robustness of CFEs and introduces BetaRCE, a post-hoc method that enhances existing CFEs with robustness guarantees and user-adjustable confidence levels.

Findings

BetaRCE provides robust counterfactual explanations with probabilistic guarantees.

Experimental results show BetaRCE outperforms baselines in robustness and plausibility.

BetaRCE is adaptable to any model and change type, with easy hyperparameter tuning.

Abstract

Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.